[gmx-developers] XML

[Michael Shirts]

Hi, all-

I'm in Lyon, at the CECAM conference on open standards for molecular and
atomistic simulations.  We've just spent the last several days discussing XML
and other sorts of open standards. The consensus seems to be, talking to
people in other simulation communities (DFT, electronic structure) that XML is
the way to go.

I've talked with Michael Crowley, who is heavily involved with both CHARMM and
AMBER (he works for both David Case and Charlie Brooks) and he is also very
positive about moving going to XML, as are people involved with TINKER and
some of the smaller European programs.  What we all expecially interested in
is making a COMMON XML format for such things as trajectory files and
parameter files.  With CHARMM, AMBER, and GROMACS gravitating towards the same
format, that would help testing and comparisons immensely, and would nudge
everybody else in the same direction.

There are standard formats for binary data as well - NetCDF seems to be the
format of choice - one can imagine that an output file would be some mix of
XML and binary data.

One possiblity for input files, of course, is that either people could write
them by hand, or there could be a simple interpreting program -- either a gui
or a parser, that could hand-hold those that didn't want to / feel comfortable
writing the XML.  There are a wide variety of XML parsers already available
for various code bases.

Anyway, a working group will be forming to address the issue of standards for
macromolecular modeling, so let me know if you're interested, and I can pass
your information along to the organizers.

Cheers,
Michael Shirts
Pande Group
Department of Chemistry
Stanford University